Limits to growth (of heat pumps)

Because the temperature of the ground, a few metres down, stays sluggishly
close to 11 °C, whether it’s summer or winter, the ground is theoretically
a better place for a heat pump to grab its heat than the air, which in
midwinter may be 10 or 15 °C colder than the ground. So heat-pump advisors
encourage the choice of ground-source over air-source heat pumps,
where possible. (Heat pumps work less efficiently when there’s a big temperature
difference between the inside and outside.)

However, the ground is not a limitless source of heat. The heat has to
come from somewhere, and ground is not a very good thermal conductor.
If we suck heat too fast from the ground, the ground will become as cold as
ice, and the advantage of the ground-source heat pump will be diminished.

In Britain, the main purpose of heat pumps would be to get heat
into buildings in the winter. The ultimate source of this heat is the sun,
which replenishes heat in the ground by direct radiation and by conduction
through the air. The rate at which heat is sucked from the ground
must satisfy two constraints: it must not cause the ground’s temperature
to drop too low during the winter; and the heat sucked in the winter must
be replenished somehow during the summer. If there’s any risk that the
natural trickling of heat in the summer won’t make up for the heat removed
in the winter, then the replenishment must be driven actively – for example
by running the system in reverse in summer, putting heat down into the
ground (and thus providing air-conditioning up top).

Let’s put some numbers into this discussion. How big a piece of ground
does a ground-source heat pump need? Assume that we have a neighbourhood
with quite a high population density – say 6200 people per km2
(160 m2 per person), the density of a typical British suburb. Can everyone
use ground-source heat pumps, without using active summer replenishment?
A calculation in Chapter E (p303) gives a tentative answer of no:
if we wanted everyone in the neighbourhood to be able to pull from the
ground a heat flow of about 48 kWh/d per person (my estimate of our
typical winter heat demand), we’d end up freezing the ground in the winter.
Avoiding unreasonable cooling of the ground requires that the sucking
rate be less than 12 kWh/d per person. So if we switch to ground-source
heat pumps, we should plan to include substantial summer heat-dumping
in the design, so as to refill the ground with heat for use in the winter. This
summer heat-dumping could use heat from air-conditioning, or heat from

Figure 21.12. How close together can ground-source heat pumps be packed?
area per person (m2)
Bangalore 37
Manhattan 39
Paris 40
Chelsea 66
Tokyo 72
Moscow 97
Taipei 104
The Hague 152
San Francisco 156
Singapore 156
Cambridge MA 164
Sydney 174
Portsmouth 213
Table 21.13. Some urban areas per person.